SAVRN Model Hub
AI Training Datasets
Each dataset with its card, its structure (every split, row count and column type), its files, its license, and the models that disclose training on it.
11,122Models
3,164Datasets
285Papers
6,792Publishers
11,852Sourced relationships
Updated 2026-10-18 · How the library is built
3,164 datasets, sorted by most downloaded.
Monthly preprocessed limit order book training-data archives, January 2022 through February 2026. Files retain their original SquashFS format. Access to data requires approval from the dataset owner. The repository includes an index, backup metadata, and SHA-256 manifests. After downloading, verify integrity using sha256sum -c SHA256SUMS.txt. SquashFS images can be inspected or extracted with SquashFS tools. No redistribution license is granted by this dataset card.
The instance data-lab-demo contains example datasets generated using the latest version of instance Data Lab. The datasets are summarized in the binding-target-cards and specificity-target-cards data quality reports. For a more thorough explanation, refer to the Instance Data Lab report. - affinityscores contain an early preview of an upcoming product release that combines bindings scores from multiple doses into a single affinity score that's consistently scaled across datasets.
The frozen evaluation sets for the PixelProof paper. Each eligible replay-verified world contributes exactly five deterministic, pixel-distinct instances. The five instances of a world share its selected question text while varying the rendered scene and the gold answer, which the world's inverse program recovers from the image alone. The config names, the section column, and the ledgers/section files are holds spatially-steered worlds. The exclusions are the worlds that passed the 200-scene replay (1,301 profile-steered and 875 spatially-steered) but produced fewer than five pixel-distinct scenes; each one is documented in the corresponding ledgers/sectionexclusions.csv file. Each row…
测试内容已于 2026-10-15 清理，当前尚无已发布数据。metadata 中 task、trajectory、frame 均为 0；已知的数据说明保留，未知版本、联系方式、参考资料等留空。初始化不代表正式发布或质量通过。 按机器人区分子数据集，压缩后的完整 Zarr 保存到对应目录，不在仓库根目录上传裸 Zarr 或任务压缩包；目录在首次正式上传时创建。 分批默认目标约 20 GiB，以完整 episode 为边界。仅在人工审核并显式点击上传后发布；task 按任务简称去重，trajectory 和 frame 根据实际已发布 episode 更新。 官方检查输出与 humanreview 保存于 MTEAI-QC。真实标注图、机构及传感器图片、封面视频等正式提交材料后续补充。
This snapshot contains 55,050 distinct normalized text bodies and 65,209 source records. Generated 2026-10-02T03:18:37.852690+00:00. One Parquet row represents one normalized source body, with original extracted text, source title, URL, extraction quality, observed retrieval time, rule version, topic labels and all current-body source aliases. Duplicate URLs/editions are preserved as provenance, not counted as additional distinct documents. The canonical record is the lexicographically smallest eligible document identifier. Topic labels overlap. Selection requires an eligible assessment whose body checksum still matches the current document, with fulltext or ocrtext extraction. This is…
Model Collections
Hand-picked starting points, each with the reason it exists.
Collection · 4 entries
Embedding models for retrieval
Sentence and document embedding models used to build retrieval systems. Dimension and sequence length matter more than size here, and both come from the publisher.
Collection · 4 entries
Models that fit on one accelerator
Models whose publisher-reported parameter count puts them within reach of a single accelerator at common precisions. Memory needed depends on precision and serving configuration, so treat the parameter count as the starting point, not the answer.
Collection · 6 entries
Open-weight text models worth knowing
Widely used open-weight language models, chosen because each one is a distinct family rather than a variant of the one above it. Selection, not a ranking.
Collection · 3 entries
Speech and audio models
Recognition and synthesis models, grouped so the two directions are easy to compare.
Related SAVRN Research
The hub sits beside SAVRN's market data and infrastructure research: what models cost to run, and what it takes to run them.
SAVRN Index
What open models cost to run
The same open-weight model priced by every host that serves it, per million tokens.
Research Hub
Data center trackers and maps
Moratoriums, permits, power, water and capital behind the facilities that run these models.
Method
How the Model Hub is built
Sources, evidence labels, refresh behaviour, and the limits of every comparison here.